The Age Of Video: How Mobile Networks Must Evolve
In 3G, mobile network operators find themselves in a situation where their core network is composed of many complex elements (GGSN, EPC, browsing gateways, proxies, DPI, PCRF…) that are extremely specialized but have been designed with transactional data in mind. Radio access is a scarce resource, with many operators battling with their regulators to obtain more spectrum. The current model to purchase capacity, based on purchasing more base stations, densifying the network is finding its limits. Costs for network build up are even expected to exceed data revenues in the coming years.
On the technical front, some operators are hitting the Shannon’s law, the theoretical limit for spectrum efficiency. Diminishing returns are the rule rather than the exception as the RAN (Radio Access Network) becomes denser for the same available spectrum. Noise and interferences increase.
On the financial front, should an operator follow the demand, it would have to double its mobile data capacity on a yearly basis. The projected revenue increase for data services shows only a CAGR of 20% through 2015. How can operators keep running their business profitably?
Operationally, doubling capacity every year seems impossible for most networks who look at 3 to 5 years roll out plans. A change of paradigm is necessary.
Solutions exist and start to emerge. Upgrade to HSPA +, LTE, use smaller cells, changing drastically the pricing structure of the video and social services, network and video optimization, offload part of the traffic to wifi, implement adaptive bit rate, optimize the radio link, cache, use CDNs, imagine new business models with content providers, device manufacturers and operators…
The main issue is one of network intelligence. Mobile network operators want their network utilization optimized, not minimized. Traffic patterns need to be collected, analyzed, represented so that data and particularly video can be projected, but not at the country, multi-year level as of today. It is necessary to build granular network planning capacity per sector, cell at RAN, Core and Backhaul levels with tools that are video aware. Current DPI and RAN monitoring tools cannot detect video efficiently and analyze it deeply enough to allow for pattern recognition. Additionally, it is necessary to be able to isolate, follow and act on individual video streams on a per subscriber, per service, per property, per CDN level, not simply at the protocol level.
Current mobile network analytics capabilities are mostly inherited from 3G. DPI and traffic management engines rely mostly on protocol analysis and packet categorization to perform their classification and reporting. Unfortunately, in the case of video, this is insufficient. Video takes many forms in mobile networks and is delivered over many protocols (RTSP, RTMP, HTTP, MPEG2TS…). Recognizing these protocols is not enough to be able to perform the necessary next steps. Increasingly, video traffic is delivered over HTTP progressive download. Most current analytics capabilities cannot recognize video as a traffic type today. They rely on url recognition rather than traffic analysis. This leads to issues: how do you differentiate when a user is browsing between YouTube pages from when he is watching a video? How do you discriminate embedded videos in pages? How do you recognize You Tube embedded videos in Facebook? How do you know whether a video is an advertisement or a main programming? How do you know whether a video should be delivered in HD or lower resolution?
It is necessary, in order to categorize and manage video accurately to recognize the video protocol, container, codec, encoding rate, resolution, duration, origin at the minimum to be able to perform pattern recognition.
Measure Experience, not Speed or Size
The next necessary step after identifying and indexing the video traffic is the capacity to grade it from a quality standpoint. As video quality becomes synonymous to network quality in viewers’ mind, mobile network operators must be able to measure and control video quality. Current capabilities in this space are focused on measuring network speed and content size and inferring user satisfaction. This is inadequate
Any hope of monetizing mobile video for mobile network operators beyond byte accounting relies on being able to reliably grade video content in term of quality. This quality measurement is the cornerstone to provide subscribers with the assurance that the content they view is conform to the level of quality they are entitled to. It is also necessary for network operators to establish baseline with content providers and aggregators who view content quality as one of the main elements of pricing.
A uniform Quality of Experience (QoE) measurement standard is necessary for the industry to progress. Today, there is no valid QoE metric for mobile networks, leaving mobile operators relying on sparse proprietary tools, often derived or created for broadcast and professional video, wholly inadequate for mobile networks. Mobile network operators must be able to measure the QoE per video, subscriber, session, sector, cell, origin, CDN if they want to create intelligent charging models.
Analyze, Segment Consumers and Traffic
Mobile network operators have been segmenting efficiently their customer base, building packages, bundles and price plans adapted to their targets. In the era of video, it is not enough.
Once traffic is identified, indexed, recognized, it is important to segment the population and usage. Is video traffic mostly from premium content providers and aggregators or from free user generated sites? Are videos watched mostly long form or short form? Are they watched on tablets or smartphones? Are they very viral and watched many times or are consumers more following the long tail? All these data and many others are necessary to understand the nature of subscribers’ consumption and will dictate the solutions that are most appropriate. This is a crucial step to be able to control the video traffic.
Once video traffic is correctly identified and indexed, it becomes possible to manage it. It is a controversial topic as net neutrality as a concept is far from being settled, at least in the mobile world. My view is that in a model were scarcity (spectrum, bandwidth) and costs are borne by one player (operators) while revenue and demand are borne by others (content providers and subscribers), net neutrality is impractical and anti-competitive. Unlike in fixed network, where quasi-unlimited capacity and low entry costs allow easy introduction of content and services, mobile networks’ cost structures and business models are managed systems where demand outgrows capacity and therefore negate equal access to resources. For instance, no one is talking about net neutrality in the context of television. I believe that operators will be able to discriminate traffic and offer models based on subscribers and traffic differentiation, many already can. It is just a recognition that today, with current setup, traffic gets degraded naturally as demand grows and DPI and traffic management engine are already providing means to shape and direct traffic to everyone’s best interest. No one could think of networks where P2P file sharing traffic could go unchecked and monopolize the network capacity.
Additionally, all videos are not created equal. There are different definitions, sizes, encoding rates. There are different qualities. Some are produced professionally, with big budgets, some are user generated. Some are live, some are file based. Some are downloaded, some are streamed. Some are premium, some are sponsored, some are freemium, some are free… Videos in their diversity bear the key to monetization.
The diversity of videos and their mode of consumption (some prefer to watch HD content in the highest quality, and will prefer download over streaming, others prefer a video that runs uninterrupted, with small load time even with a lesser quality…) is the key to monetization.
Mobile network operators must be able to act based on video and subscribers attribute and influence the users’ experience. Being able to divert traffic to other bearers (LTE, Wifi…), to adjust a video quality on the fly are important steps towards creating class of services, not only amongst subscribers but also between content providers.
It is important as well to enable subscribers to select specific quality levels on the fly and to develop the charging tools to provide instant QoE upgrades.
With the capacity to detect, measure, analyze, segment, control and manage, operators can then monetize video. The steps highlighted here provide means for operators to create sophisticated charging models, whereby subscribers, content providers and aggregators are now included in a virtuous value circle.
Operators should explore creating different quality threshold for the video content that transits through their network. It becomes a means to charge subscribers and / or content providers for premium guaranteed quality.